'New Audio Op Amp - OPA1622' or Another Successful Headphone Amp Project!!

0 Members and 1 Guest are viewing this topic. Read 2769 times.


  • Full Member
  • Posts: 1574

I had two more small head amp projects in the hopper, and I finally got around to complete one.

There is a fellow in Paris, France that saw the TI just designed a new very small op amp specifically for headphone amps and had
to try it out!

TI has design circuits, test circuits etc.."Chris" designed a circuit and layed out a PCB and I wanted to hear this new headphone amp chip the OPA 1622.

Heres the link to the TI datasheet: http://www.ti.com/product/OPA1622/datasheet

The dimensions of this 10 pin chip: OPA1622   VSON (10)   3.00 mm × 3.00 mm

This presents a problem for us old school solder DIY'ers!

It made me buy a hot air gun station, learn how to use solder paste and do micro soldering !!!
It was at first intimidating, but after some practice and sitting on the project for a few weeks I started!!

The OPA1622 is U4 all hot air gunned down....notice the 2mm x 2mm IC on yet soldered down!

All 5 of the IC's a vregulator 7805 are installed.

Here is the rest of the goodies all nice and neat ready for testing!

What it looks like packaged up less a volume knob:


Here are Chris's design comments:

Like some of you I followed with great interest the thread 'New Audio Op Amp - OPA1622' by TI’s Johnnc124. I wasn’t initially specifically interested by headamps but as the OPA1622 was simple to implement and was supposed to provide good results I decided to setup a project with it. More than building an headamp this project was aimed to confirm I could achieve designing a PCB with a CAD software and also soldering SMD devices, including ones in DFN or QFN package.

I’have finally been through all steps, I’m happy with the result (OPA1622 sounds very good !) and I’m now confident to move to projects that will require complicated PCB and good soldering capabilities.
Now, I have to finish assembly and testing of a Salas DGC3 in order to do some comparisons!

The main requirements for AmpCasq where :
It should stand in a small aluminum box, in order to sit on desk
It should be powered by an AC adapter I had, which provided 17VDC
It should have a volume control
It should have RCA and jack 3.5 inputs
It should have jack 3.5 & 6.35 outputs
It should allow me to experiment crossfeed
Final price should stay under control

About the schematics:
Ampcasq was designed in September 2016, at this time LM27762 was announced but not available. So, I did stay with LM27761 for -5V and LP38798 for +5V; LP5907 was a bit short on current. The 7805 lifted by R17/R12 was added to limit LM27761 input voltage to 5.5V, its max value. 7805 allow Ampcasq to be powered by any DC source from 10V to 35V, 300mA max.
Power Input:
D1 to protect against bad connection.
D2, R23, R10, R11, C24 to manage the Enable pin of the OPA. Objective was to avoid noise in the can at power up/down. This partly works: mute the output when enable Low (<~0.8V) but can’t avoid noise in can when power supply capacitors are discharching and OPA1622's supply getting out of tolerance. On power off with SW1, OPA1622 is muted as C24 is quickly discharged through R23, D2 and SW1. D2 is a Schottky, not for speed but for low drop (Vf close to 0.3V).
It is based on an Xen-Audio (EUVL) paper: Xen Audio cross feed.
There’re three positions for the switch: Low Xfeed – No Xfeed – High Xfeed. Low Xfeed is Ok, High is exaggerate: there’s too much bass boost. After some listening sessions I finally do not use the Xfeed.
Gain + out
Not much to say.


I will do much listening and compare to my other amps and projects...so far it worked the first time powered up!!
This is always a good thing!

All the best


  • Full Member
  • Posts: 3387
  • Science is not a democracy - Earl Geddes
    • 2 channel/7 channel setup
Very nicely done! Like you, I have been knee deep in SMD projects as of late so I can relate! It's not that hard, takes a steady hand, patience, LOTS of flux, tweezers, a 1.6mm tip, plenty of LED lighting, a 10X Loupe or Flip Focus lens and voila, a whole new world awaits.

Some measurements of the circuit described would be nice...

Thanks for sharing!



  • Full Member
  • Posts: 1574
Hey Anand...

I just re-did my garage space and made it super clean, a new workbench and LED lighting...the head magnifier, hot air gun, solder paste, flux etc...its really neat to watch the paste change states and parts jiggle into place!! LOL.

I dont have any measurements on my amp but here are the specs for the OPA 1622:

High-Fidelity Sound Quality
Ultralow Noise: 2.8 nV/√Hz at 1 kHz
Ultralow Total Harmonic Distortion + Noise:
–119 dB THD+N (142 mW/Ch into 32 Ω/Ch)
Wide Gain Bandwidth Product:
32 MHz (G = +1000)
High Slew Rate: 10 V/μs
High Capacitive-Load Drive Capability: > 600 pF
High Open-Loop Gain: 136 dB (600-Ω Load)
Low Quiescent Current: 2.6 mA per Channel
Low-Power Shutdown Mode With Reduced Pop and Click Noise: 5 μA per Channel
Short-Circuit Protection
Wide Supply Range: ±2 V to ±18 V

The OPA1622, dual, bipolar-input, SoundPlus™ audio operational amplifier achieves a very low,
2.8-nV/√Hz noise density with an ultralow THD+N of –119.2 dB at 1 kHz and drives a 32-Ω load at 100-mW output power. The OPA1622 offers extremely-high ac PSRR and CMRR specifications that eliminate noise from power supplies, making the OPA1622 ideal for use in portable-audio applications. This device also has a high output-drive capability of +145 mA per –130 mA.

The OPA1622 operates over a very wide supply range of ±2 V to ±18 V, on only 2.6 mA of supply current per channel. The OPA1622 op amp is unity-gain stable and provides excellent dynamic behavior over a wide range of load conditions. The OPA1622 includes a shutdown mode, allowing the amplifiers to be switched from normal operation to a standby current that is typically less than 5 µA. This shutdown feature is specifically designed to eliminate click and pop noise when transitioning into or out of shutdown mode.

Chris who designed this circuit did state: Based on my measurements on my board with the +/- 5V supply, OPA 1622 can drive up to 3Vpeak on 32R//1nF load with no 'visible' distortion (I didn't have FFT capabilities for now) which is consistent with datasheet.

I have a friend in Texas that is going to get a FFT capable test equip and when that happens I will send this to him for measurments.

I have compared to a O2 Booster Board amp with it and they are really almost identical is sound.



  • Full Member
  • Posts: 3387
  • Science is not a democracy - Earl Geddes
    • 2 channel/7 channel setup

Thank you most graciously for the summary. I'll look forward to the FFT, THD, IMD, etc... when available.



  • Full Member
  • Posts: 1574
I asked Chris for any detail specs he has:

Using the headphone power calculator of Bob Robinette give the following results with the cans you mentioned :
•       DT1350 (80R, 109dB/mW), OPA1622 with +/-5V supply should be ok, it even can reach the max SPL of the can : 129dB
•       T1 (600R, 102dB/mW), max SPL with OPA1622 with +/-5V  (clipping at 4.1V) is 113dB. Certainly too short, it should be better with +/- 10 V supply, allowing 120dB SPL max.
•       T90 (250R, 102dB/mW), OPA1622 with +/-5V supply should be ok, max reachable SPL : 117dB
The max SPL section of the spreadsheet (‘How Loud Will an Amp Drive Your Headphones’) is based on the max output voltage of the OPA1622 which is appx Rail-0.9V with most of loads.
But those are only figures…listening session will confirm if AmpCasq fit the various cans



  • Full Member
  • Posts: 1574

Chris the designer of this amp has lent his am out to a person near Paris who will do some measurments and I will post as soon
as I have them.

Also we did some modification over the original design...the X-Feed circuitry did absolutely nothing for the audio experience and the gain for lower impedance headphones (80 ohms or less) was a little high and the volume control could not be opened up much more than the 9 o'clock positon...so out went x-feed and we used the three position x-feed switch to allow for three selectable gain positions, 3db, 6db and 9db. Had to change a bunch of resistors...but it works much better over a wider set of impedances.